The WORC database consists in total of 930 patients composed of six datasets gathered at the Erasmus MC, consisting of patients with: 1) well-differentiated liposarcoma or lipoma (115 patients); 2) desmoid-type fibromatosis or extremity soft-tissue sarcomas (203 patients); 3) primary solid liver tumors, either malignant (hepatocellular carcinoma or intrahepatic cholangiocarcinoma) or benign (hepatocellular adenoma or focal nodular hyperplasia) (186 patients); 4) gastrointestinal stromal tumors (GISTs) and intra-abdominal gastrointestinal tumors radiologically resembling GISTs (246 patients); 5) colorectal liver metastases (77 patients); and 6) lung metastases of metastatic melanoma (103 patients). For each patient, either a magnetic resonance imaging (MRI) or computed tomography (CT) scan, collected from routine clinical care, one or multiple (semi-)automatic lesion segmentations, and ground truth labels from a gold standard (e.g., pathologically proven) are available. All datasets are multicenter imaging datasets, as patients referred to our institute often received imaging at their referring hospital. The dataset can be used to validate or develop radiomics methods, i.e., using machine or deep learning to relate the visual appearance to the ground truth labels, and automatic segmentation methods. See also the research article related to this dataset: Starmans et al., Reproducible radiomics through automated machine learning validated on twelve clinical applications, Medical Image Analysis, Submitted.

Radiomics uses quantitative medical imaging features to predict clinical outcomes. While many radiomics methods have been described in the literature, these are generally designed for a single application. The aim of this study is to generalize radiomics across applications by proposing a framework to automatically construct and optimize the radiomics workflow per application. To this end, we formulate radiomics as a modular workflow, consisting of several components: image and segmentation preprocessing, feature extraction, feature and sample preprocessing, and machine learning. For each component, a collection of common algorithms is included. To optimize the workflow per application, we employ automated machine learning using a random search and ensembling. We evaluate our method in twelve different clinical applications, resulting in the following area under the curves: 1) liposarcoma (0.83); 2) desmoid-type fibromatosis (0.82); 3) primary liver tumors (0.81); 4) gastrointestinal stromal tumors (0.77); 5) colorectal liver metastases (0.68); 6) melanoma metastases (0.51); 7) hepatocellular carcinoma (0.75); 8) mesenteric fibrosis (0.81); 9) prostate cancer (0.72); 10) glioma (0.70); 11) Alzheimer's disease (0.87); and 12) head and neck cancer (0.84). Concluding, our method fully automatically constructs and optimizes the radiomics workflow, thereby streamlining the search for radiomics biomarkers in new applications. To facilitate reproducibility and future research, we publicly release six datasets, the software implementation of our framework (open-source), and the code to reproduce this study.

Patients with BRAF mutated (BRAF-mt) metastatic melanoma benefit significantly from treatment with BRAF inhibitors. Currently, the BRAF status is determined on archival tumor tissue or on fresh tumor tissue from an invasive biopsy. The aim of this study was to evaluate whether radiomics can predict the BRAF status in a non-invasive manner. Patients with melanoma lung metastases, known BRAF status, and a pretreatment computed tomography scan were included. After semi-automatic annotation of the lung lesions (maximum two per patient), 540 radiomics features were extracted. A chest radiologist scored all segmented lung lesions according to the Lung Image Database Consortium (LIDC) criteria. Univariate analysis was performed to assess the predictive value of each feature for BRAF mutation status. A combination of various machine learning methods was used to develop BRAF decision models based on the radiomics features and LIDC criteria. A total of 169 lung lesions from 103 patients (51 BRAF-mt; 52 BRAF wild type) were included. There were no features with a significant discriminative value in the univariate analysis. Models based on radiomics features and LIDC criteria both performed as poorly as guessing. Hence, the BRAF mutation status in melanoma lung metastases cannot be predicted using radiomics features or visually scored LIDC criteria.

Neuroinflammation has been implicated in frontotemporal lobar degeneration (FTLD) pathophysiology, including in genetic forms with microtubule‐associated protein tau (MAPT) mutations (FTLD‐MAPT) or chromosome 9 open reading frame 72 (C9orf72) repeat expansions (FTLD‐C9orf72). Iron accumulation as a marker of neuroinflammation has, however, been understudied in genetic FTLD to date. To investigate the occurrence of cortical iron accumulation in FTLD‐MAPT and FTLD‐C9orf72, iron histopathology was performed on the frontal and temporal cortex of 22 cases (11 FTLD‐MAPT and 11 FTLD‐C9orf72). We studied patterns of cortical iron accumulation and its colocalization with the corresponding underlying pathologies (tau and TDP‐43), brain cells (microglia and astrocytes), and myelination. Further, with ultrahigh field ex vivo MRI on a subset (four FTLD‐MAPT and two FTLD‐C9orf72), we examined the sensitivity of T2*‐weighted MRI for iron in FTLD. Histopathology showed that cortical iron accumulation occurs in both FTLD‐MAPT and FTLD‐C9orf72 in frontal and temporal cortices, characterized by a diffuse mid‐cortical iron‐rich band, and by a superficial cortical iron band in some cases. Cortical iron accumulation was associated with the severity of proteinopathy (tau or TDP‐43) and neuronal degeneration, in part with clinical severity, and with the presence of activated microglia, reactive astrocytes and myelin loss. Ultra‐high field T2*‐weighted MRI showed a good correspondence between hypointense changes on MRI and cortical iron observed on histology. We conclude that iron accumulation is a feature of both FTLD‐MAPT and FTLD‐C9orf72 and is associated with pathological severity. Therefore, in vivo iron imaging using T2*‐weighted MRI or quantitative susceptibility mapping may potentially be used as a noninvasive imaging marker to localize pathology in FTLD.

Early pathological features of frontotemporal lobar degeneration (FTLD) due to MAPT pathogenic variants (FTLD-MAPT) are understudied, since early-stage tissue is rarely available. Here, we report unique pathological data from three presymptomatic/early-stage MAPT variant carriers (FTLD Clinical Dementia Rating [FTLD-CDR] = 0–1). We examined neuronal degeneration semi-quantitatively and digitally quantified tau burden in 18 grey matter (9 cortical, 9 subcortical) and 13 white matter (9 cortical, 4 subcortical) regions. We compared presymptomatic/early-stage pathology to an intermediate/end-stage cohort (FTLD-CDR = 2–3) with the same variants (2 L315R, 10 P301L, 6 G272V), and developed a clinicopathological staging model for P301L and G272V variants. The 68-year-old presymptomatic L315R carrier (FTLD-CDR = 0) had limited tau burden morphologically similar to L315R end-stage carriers in middle frontal, antero-inferior temporal, amygdala, (para-)hippocampus and striatum, along with age-related Alzheimer’s disease neuropathological change. The 59-year-old prodromal P301L carrier (FTLD-CDR = 0.5) had highest tau burden in anterior cingulate, anterior temporal, middle/superior frontal, and fronto-insular cortex, and amygdala. The 45-year-old early-stage G272V carrier (FTLD-CDR = 1) had highest tau burden in superior frontal and anterior cingulate cortex, subiculum and CA1. The severity and distribution of tau burden showed some regional variability between variants at presymptomatic/early-stage, while neuronal degeneration, mild-to-moderate, was similarly distributed in frontotemporal regions. Early-stage tau burden and neuronal degeneration were both less severe than in intermediate-/end-stage cases. In a subset of regions (10 GM, 8 WM) used for clinicopathological staging, clinical severity correlated strongly with neuronal degeneration (rho = 0.72, p < 0.001), less strongly with GM tau burden (rho = 0.57, p = 0.006), and did not with WM tau burden (p = 0.9). Clinicopathological staging showed variant-specific patterns of early tau pathology and progression across stages. These unique data demonstrate that tau pathology and neuronal degeneration are present already at the presymptomatic/early-stage of FTLD-MAPT, though less severely compared to intermediate/end-stage disease. Moreover, early pathological patterns, especially of tau burden, differ partly between specific MAPT variants.